How Much Information

How much information? Chapters from the history of the accounting of information flows and knowledge stocks István Dienes - scientific advisor Cég-INFO Ltd., Budapest [email protected], http://free.x3.hu/infostat

"How much information?" A question, which has been raised by electric engineers, mathematicians and the researchers studying the "information revolution" and "information society". In the present paper I outline the history of information accounting and provide an assessment of strengths and weaknesses of various approaches.

1. Chapters from the history of information accounting

Chapter 1. Pioneers’ epoch

It was Ithiel de Sola Pool and Inose, who were pioneering in the attempts to account information flows and knowledge stocks in a country, particularly in the U.S. and Japan [Pool 1984]. They introduced "production" and "consumption" of information and determined the value of these variables. As a unit of their measurements "word" was introduced. Empirical - and arbitrary - rules to express the the volume of various carriers in a common unit have been defined. While they accounted various carriers and carriers altogether, their tables can not be viewed as accounts as defined in book-keeping. Marc Uri Porat in his unical treatise on U.S. information economy has first provided a comrehensive economi view of all information activities of a country.

Chapter 2 System of National Information Accounts - an effort in Budapest

Lakoff's epochal works on metaphors [Lakoff] have revealed that language reflects "folk-psychology" of phenomena and processes: the world itself, as it is represented in our mind and brain. In English, Hungarian and in a number of other languages, information is a fluid; it has volume, it flows, can be produced, outputted, inputted or consumed, and it can be stored. Commodities and value, in an economic sense, also obey the rules governing the behavior of "fluids", this is why micro-, and macroeconomic accounts can be defined and applied worldwide both in book-keeping and economy.

To explain, forecast and understand the behavior of fluids in a particular situation, the generic theory, the equations of heat and mass transfer should be adopted and applied. The theories concerning economic behavior of "information" that predict or declare that information can be consumed unless its volume in the process would change, contradict to the intuitive contents of the concept itself, which reflects in the word usage, and the traditional official statistics of information goods.

An exact theory should be based upon solid concepts, operational measurement procedures in harmony with theory of transport processes. A standard system of standard accounts is needed to compare information-related phenomena in the countries of the world. In any particular measurement the conditions of the situation should be exactly determined and extrinsic quantities should be measured and introduced.

It was the Japanese Planning Agency, which first has published bit-based figures for the "production" of information in Japan. Figures for Hungary, which have been compiled by the Poo1's method have been published by the present author: [Dienes92]. At the same time, in 1990 and 1991 the figures and methodology of information accounting have been reviewed and reconsidered. As a result, a new methodology, called SNIA, System of National Information Accounts, was developed, which is directly related to SNA's approach and concepts. Then, information balances of Hungary for the period 1945-1990 have been compiled [Dienes92].The political, practical and statistical considerations behind SNIA, as well as figures for Hungary, can be viewed at http//free.x3.hu/infostat.

2.1 Subjects and objects of the accounting in the SNIA Due to the new approach, any account of information should be in accordance with national accounts, both concerning concepts and figures. SNIA has been based upon the same fundamental general concepts as [SNA93]. These include actor, institutional unit, sector, good, service, commodity, economic transaction, economic stock, economic flow, account, balance. SNIA covers those subjects, objects, relations, acts, actions and activities which are subject to legal definition, require comprehensive treatment and can be subject to operationalization.

The vary same concepts, however, have sometimes been interpreted in a wider environment to include events and objects that are beyond the present scopes of SNA, but seem to be important, and new concepts also were introduced. New integrated sectors; the main groups of social actors of information transactions, which are relevant to policy making, should also be added to those of SNA, to reflect real situation, phenomena and processes. SNIA classifies transactions by commodity.

Information is understood here as something which forms or formed within (the brain of) either human or machine actors, or is represented in/on the goods/services outputted. Volume/amount of information carried by physically existing goods and services - in a standardized communication situation as defined here - is an attribute of goods and services which is very similar to extrinsic physical parameters. This interpretation seems to be in harmony with exact theories of information, such as Shannon's.

The fundamental question of SNA is "Who does What by What means for What purpose with Whom in exchange for What with What changes in Stocks?"

In accordance with this, an action is called an information flow if it leads to the change of information assets of participants. A transaction can be called an information transaction, if it implies the flow of information. There is a class of goods and services whose main function is just to convey/carry information, these are called information goods and services. Information goods do not include machines like computers, office and telecommunications devices; rather they include books, diskettes, records and a number of other durable media.

The non-exhaustive list of main categories, in terms of which SNIA describes information phenomena, includes information-input, -output, -production (gross and information added), -consumption, -use, -asset, -accumulation, -capital, -stock, productive -consumption, -capital consumption, -export, -import and externalities. Unlike SNA, SNIA makes difference between use and consumption: while consumption assumes the annihilation of the good or service, its use does not. Almost the whole arsenal of macroeconomics is expected to be transformable to SNIA.

The most outstanding difference between SNA and SNIA is in the way they valuate flows and stocks. While valuation of transactions in SNA has practically been based upon general substitutability for money, opportunity of exchange for money, SNIA provides the opportunity of vaulation based upon general substitutability of information-commodities for a digital record, opportunity of recording, exchange for a digital record. Accordingly, it is not information in SNIA, which is considered as resource or product, but goods (including non-durable signals) which carry/convey information. This is in agreement with SNA, where it is not "value" which is considered as a resource or product, but goods which have or carry value. Information interpreted in such manner, can be treated similarly to energy. Various versions of SNIA accounts involve the information that goods and services carry at "sensory" or at "perception" level.

As an illustration of the several tables and figures, the Figure 1. shows output, consumption, of information commodities in Hungary in the period 1975-1990.

So far, basic SNIA tables have been compiled in bit units. Most tables, particularly those concerning all information goods and services altogether, may be, and some actually were compiled in value units, either. These tables -- called twin tables -- show "information economy" in a commodity approach. These tables do not belong to the standard set of tables of SNA, for they contain figures for such a group of goods and services whose elements are scattered in a number of industries, group of kinds and group of services of standard SNA. This set of twin tables, however, may create a bridge, a direct linkage to SNA. Different interpretation of production, different asset and national boundaries make the correction of national SNA aggregates necessary.

While SNIA has been designed for macro level analyses, its concepts and structure make it capable to treat other communication issues and situations either; as regional communication or communications among various communities and groups. These accounts do not organically belong to the SNIA as suggested.

Chapter 3. A Standard Sytem of National Information Accounts - Budapest in Berkeley

In the frames of a Fulbright scholarship at UC Berkeley, the present author has developed the Version 1.0 of SSNIA, a standard system of national information accounts, whose inputs are standard tables of international official statistics or statistics of international organisations. The objective of the system is to provide a comprehensive, multipurpose intellectual framework and vehicle (accounts, balance sheets, tables, based on a set of internationally agreed concepts, definitions, conventions, classifications and accounting rules) which is compatible with SNA. Hundreds of standard indicators are available in the system for the purposes of description and analysis. The framework should allow the governments, business groups and citizens to analyze and display the important and relevant individual features of the countries and to provide information for the government to conduct a comprehensive -- non- industry and non-department level -- national and international information policy . In addition, it should be useful for conducting scientific research. The first two introductory chapters of SSNIA can be read at the website referred above.

Chapter 4 USSNIA - The estimation of the main SSNIA Indicators for the U.S.

The objectives of this study - conducted in Berkeley by the present author - were to - outline the "information household", the process of social reproduction of information in the United States in the extended framework of SNIA to - position the system of social reproduction of informationin a space defined by the antinomies of social reproduction: market--non-market, corporations--governments--households, production--consumption, domestic--foreign, - determine orders of magnitude of indicators, - determine the main relations among them, - identify the main processes and trends in the U.S. in the period between 1970-1990.

Furhermore I wanted to - gather experience with SSNIA trying to understand the way of operation of the system and - make statements on the feasibility of the system.

The first comparisons with De Sola Pool’s figures justified that his method – based upon words - underestimated volumes of information carried by spectacular carriers like TV or culture and entrtainment.

The phenomenon “Electronic mass media” is really an output/consumption chain: Authors produce original music, film, entertainers play, recorders record, producers multiply records, program providers replay, transmitters transmit, broadcasters broadcast, set owners display or present, viewers when viewing reproduce information. Each player productively cosnumes the output of the former. While this sequence has been analysed and followed, [Pool] never tried to follow the media processes in details. This explains the differences in quantities of machine consumable information.

Table 1. Comparison of distribution of Output in USSNIA and “supply” in [Pool] and intermediate consumption in USNIA and “consumption” in [Pool] by commodities

Output Output "Supply" 1980, Intermediate Intermediat "Consumption" 1990 % 1980 % De Sola Pool % consumption, 1990 consumptione 1980 De Sola Pool , Products exabit exabit words % exabit 1980, %exabit % words Services 99,99 99,99 97,69 100,00 100,00 82,90 Human consumable 30,33 38,86 0,07 92,33 94,25 11,07 Education 7,38 9,45 0,03 43,91 45,49 6,36 Personal oral communication s 11,11 14,23 11,12 10,11 TV shows 4,60 5,89 5,45 5,65 Supplying radio programs 0,01 0,01 0,01 0,01 Culture and entertainment 2,14 2,74 0,02 1,35 1,40 0,44 Writing 0,00 0,00 0,00 0,00 Reading 5,10 6,54 30,49 31,59 Phone 0,00 0,00 0,02 0,00 0,00 4,27 Machine 69,66 61,14 97,62 7,67 5,75 71,83 consumable TV broadcasting 20,29 18,27 25,30 3,33 4,58 51,70 Radio broadcasting 4,24 4,60 71,80 0,03 0,03 18,80 Cable TV 45,13 38,26 0,52 4,31 1,14 0,76 TV programming 0,00 0,00 0,00 0,00 Radio programming 0,00 0,00 0,00 0,00 Phone and data services 0,00 0,00 0,00 0,00 0,00 0,57 Goods 0,01 0,01 2,76 0,00 0,00 16,18 Human consumable 0,01 0,01 2,24 0,00 0,00 16,10 Paper-based 0,01 0,01 2,24 0,00 0,00 16,10 Other 0,00 0,00 0,00 0,00 Machine consumable 0,01 0,00 0,52 0,00 0,00 0,08 Videocassette 0,01 0,00 0,00 0,00 Audiocassettes 0,00 0,00 0,00 0,00 Magnetic tapes 0,00 0,00 0,00 0,00 Magnetic diskettes 0,00 0,00 0,00 0,00 Hard disks 0,00 0,00 0,00 0,00 Film 0,00 0,00 0,00 0,00 Other 0,00 0,00 0,00 0,00 Altogether 100,00 100,00 100,45 100,00 100,00 99,08 of this SNA accounted 77,04 70,59 52,93 52,64 Non-SNA accounted 22,96 29,41 47,07 47,36 Human knowledge 0,00 0,00

However, the one-semester, which the one-semester opportunity in the frames of the Fulbright scholarship, was obviously too short for such a huge a work. The main conclusions were summarized in the Spring of 1994 in the course of a number of presentations held in the U.S., including Berkeley, UCLA, UCSD and the University of Hawaii. Phragments from these presentation can be read at http://free.x3.hu/infostat. The summary also has been read at, and published in the proceedings of, the Delhi conference of IARIW.

Having returned from Berkeley I have to reorganize Hungarian official telecommunications statistics. In a three year period I desigend 14 new surveys while nthe number of TC services providers went up from one to thousands. I also had to compile a book on Hungarian information economy in the period 1992-96, and simply had no time to complete the study of the U.S. Whereas the tables remained phragmentary, I have never published them, and the book on the work is still in manuscript. Below I present some of these tables.

Table 2 is the extended information goods and services account of the United States for the year 1990. This account shows the change of the net volume of information at the beginning of the year in the five main sectors by main groups of commodities, as a consequence of output, export and negative externalities, import and positive externalities, and consumption. The corporations sector dominates information households. Information services, like services usually, can not be reserved or maintaned, hence their volume (stock) is zero. The volume of information carried by information goods is much less, then that carried by information services. More machine consumable information was outputted, than human consumable.

Table 2. Extended information goods and services account, of the United States, 1990 (Phragment, Part I)

Products Net volume of Output by informatioas of (exabit) Jan 1st. (exabit) Sectors  Corp Govt Hholds NPI-s Total Corp Govt Hholds NPI-s Total

Services 0 0 0 0 0 1 313 519 221 280 290 924 100 ##### Human consumable 0 0 0 0 0 134 727 91 580 290 924 100 517 231 Education 0 0 0 0 0 53 297 83 086 25 183 0 161 566 Personal oral communications 0 0 0 0 0 31 725 3 525 173 402 100 208 652 TV presentations 0 0 0 0 0 11 1 10 928 0 10 940 Supplying radio programs 0 0 0 0 0 0 0 94 0 94 Culture and entertainment 0 0 0 0 0 4985 0 35152 0 40136 Writing 0 0 0 0 0 0 0 Reading 0 0 0 0 0 44 706 4 967 46 165 0 95 839 Phone 0 0 0 0 0 4 0 0 0 4 Machine consumable 0 0 0 0 0 1 178 792 129 701 0 0 1 308 492 TV broadcasting 0 0 0 0 0 260 043 121 082 0 0 381 125 Radio broadcasting 0 0 0 0 0 70 994 8 618 0 0 79 612 Cable TV 0 0 0 0 0 847 754 0 0 0 847 754 TV programming 0 0 0 0 0 0 0 0 0 0 Radio programming 0 0 0 0 0 0 0 0 0 0 Phone and data services 0 0 0 0 0 - - - - -

Goods 0 3 0 0 0 1 0 0 0 100 Human consumable 0 0 0 0 0 0 0 0 0 99 Paper-based - 0 ------99 Other ------0 Machine consumable 0 3 0 0 0 1 0 0 0 1 Videocassette - 0 - - - 0 0 0 0 0 Audiocassettes - 0 - - - 0 0 0 0 1 Magnetic tapes ------Magnetic diskettes ------Hard disks ------Film - 3 - - - - 0 - - - Other - - - - - 0 - - - -

Altogether 0,00 2,62 0,00 0,00 0,00 1 313 520 221 280 290 925 100 ##### of this SNA accounted 0 0 0 0 0 906 040 83 086 60 334 0 1 049 460 Non-SNA accounted 0 3 0 0 0 407 480 138 194 230 591 100 776 363

Human knowledge - - - - - 185 023 20 557 832 185 200 #####

Table 3. Extended information goods and services account, of the United States, 1990 (Phragment, Part II.)

Products Export and Import and extenalities evoked externalities at by (exabit) (exabit) Sectors  Corp Govt Hholds NPI-s Total Corp Govt Hholds NPI-s Total

Services 4 943 1 611 1 666 0 8 220 0 0 14 338 0 14 152 Human consumable 146 427 1 666 0 2 238 0 0 4 086 0 3 900 Education 73 427 0 0 499 0 0 1 068 0 1 068 Personal oral communications - - 1 641 0 1 641 - - 1 615 0 1 615 TV shows 0 0 25 0 25 0 0 186 0 0 Supplying radio programs 0 0 0 0 0 - - - - - Culture and entertainment 73 0 0 0 73 0 0 438 0 438 Writing 0 0 0 0 0 Reading ------780 0 780 Phone 0 0 0 0 0 0 0 0 0 0 Machine consumable 4 798 1 184 0 0 5 982 0 0 10 252 0 10 252 TV broadcasting 4 088 454 0 0 4 542 0 0 9 456 0 9 456 Radio broadcasting 710 730 0 0 1 440 0 0 796 0 796 Cable TV 0 0 0 0 0 0 0 0 0 0 TV programming 0 0 0 0 0 - - 0 0 - Radio programming 0 0 0 0 0 - - - - - Phone and data services ------

Goods 2 0 0 0 2 0 0 0 0 0 Human consumable 0 0 0 0 0 0 0 0 0 0 Paper-based ------Other ------Machine consumable 2 0 0 0 2 0 0 0 0 0 Videocassette 2 0 0 0 2 0 0 0 0 0 Audiocassettes ------Magnetic tapes ------Magnetic diskettes ------Hard disks ------Film ------Other ------

Altogether 4 945 1 611 1 666 0 8 222 0 0 14 338 0 14 152 of this SNA accounted 146 427 0 0 572 0 0 11 320 0 1 505 Non-SNA accounted 4 799 1 184 1 666 0 7 650 0 0 3 018 0 12 647

Human knowledge ------

Table 4. Extended information goods and services account, of the United States, 1990 (Phragment, Part III.)

Products Net volume of information at Consusmption the end of the by (exabit)( year Sectors  Corp Govt Hholds NPI-s Total Corp Govt Hholds NPI-s Total

Services 92 512 10 279 1 728 865 100 ###### 0 0 0 0 0 Human consumable 92 511 10 279 416 103 100 518 893 0 0 0 0 0 Education 16 077 1 786 144 271 0 162 134 0 0 0 0 0 Personal oral communications 31 725 3 525 173 376 100 208 626 0 0 0 0 0 TV shows 0 0 10 916 0 10 916 0 0 0 0 0 Supplying radio programs 0 0 94 0 94 0 0 0 0 0 Culture and entertainment 0 0 40 501 0 40 501 0 0 0 0 0 Writing 0 0 0 0 0 0 0 0 0 0 Reading 44 706 4 967 46 946 0 96 619 0 0 0 0 0 Phone 4 0 0 0 4 0 0 0 0 0 Machine consumable 0 0 1 312 762 0 1 312 763 0 0 0 0 0 TV broadcasting 0 0 386 040 0 386 040 0 0 0 0 0 Radio broadcasting 0 0 78 968 0 78 968 0 0 0 0 0 Cable TV 0 0 847 754 0 847 754 0 0 0 0 0 TV programming 0 0 0 0 0 0 0 0 0 0 Radio programming 0 0 0 0 0 0 0 0 0 0 Phone and data services 0 0 0 0 0 0 0 0 0 0

Goods 0 0 0 0 0 0 0 0 0 0 Human consumable 0 0 0 0 0 0 0 0 0 0 Paper-based ------Other ------Machine consumable 0 0 0 0 0 0 0 0 0 0 Videocassette ------Audiocassettes ------Magnetic tapes ------Magnetic diskettes ------Hard disks ------Film ------Other ------

Altogether 92 512 10 279 1 728 865 100 ###### 0 0 0 0 0 of this SNA accounted 16 077 1 786 1 457 033 0 1 050 393 0 0 0 0 0 Non-SNA accounted 76 435 8 492 271 832 100 781 263 0 0 0 0 0 Human knowledge ------

The volume of information commodities that has been outputted in the U.S., totalled 1 878 heptabit in 1990. SNA-accounted commoditiees made three quarters out of the volume of output. The output has been dominated by broadcasting. The ten-year growth rate from 1980 to 1990 was 38%. Stocks have been dominated by human knowledge: both print and digital media remain subordinate when comparing estimated volumes of knowledge stocks of human individuals. Originals are far subordinate to copies. The reader is kindly invited to study the figures deposited at the site above.

Table 5. Output and growth rate of information by groups of services and goods

Ten year growth rate Products Output, 1980 (exabit) Output, 1990 (exabit) (exabit/exabit*100) Services altogether 1 319 708 1 825 724 138 Human consumable 435 902 517 231 119 Education 121 815 161 566 133 Personal oral communications 169 076 208 652 123 TV presentation 10 017 10 940 109 Supplying radio programs 94 Culture and entertainment 40 136 Writing 0 0 Reading 94 760 95 839 101 Phone 4 4 100 Machine consumable 883 807 1 308 492 148 TV broadcasting 267 946 381 125 142 Radio broadcasting 54 820 79 612 145 Cable TV 561 040 847 754 151 TV programming 0 0 - Radio programming 0 0 - Phone and data services 0 0 - Goods altogether 99 - - Human consumable 99 - - Paper-based 99 - - Other 0 - - Machine consumable - - -

Commoditization of products is a persistent phenomenon of the history of economy.

Information transfers are uncompensated information flows. In the twenty-year period between 1970 and 1990 the share of education decreased from 26% to 14%, while the share of broadcasting increased from 46% to 61%. Socially irresponsible – mainly private - broadcasting obviously replaces socially responsible education, with all consequences: violence, agressivity, corruption of traditional values.

Table 6. Information transfers by main groups of commodities

Education Broadcasting Others Altogether Altogether Altogether per Year (exabit) (exabit) (exabit) (exabit) (Output=100) capita (petabit)

1990 108 269 460 738 184 425 753 431 41 3,01 Transfers=100 14,37 61,15 24,48 1980 106 067 322 766 143 925 572 759 43 2,52 Transfers=100 18,52 56,35 25,13 1970 108 487 190 843 115 328 414 658 50 2,02 Transfers=100 26,16 46,02 27,81

Significant volumes of information are accompanied to economic transactions as “ordering of a commodity”, “billing”, “money transfer”,etc. Beyond SNA-like indicators the volume of these information transactions has been estimated. These figures are shown in table 1.

Table 7. Intersectoral flow of information output (Exabit, phragment)

To From Corporations Governments Households NPI-s Rest of the world Altogether sources Corporations 1 237 073 0 4 943 1 313 519 Governments 212 787 0 1 611 221 280 Households 290 824 100 1 666 305 252 NPI-s 0 0 100 0 0 100 Rest of the world 0 0 14 327 0 14 142 Altogether uses 0 0 1 755 111 100 8 220 1 854 293 1 763 432

Table 8. Obligatory supply of information closely connected to transactions

Classes of examples Sorce of obligation Estimated magnitude of Source of raw data Estimated volume of order of volume of information (Exabit) information 10 to k, k= A. Information for soliciting business services A.1 Services, goods non- Interestedness specified Order A.2 Postal and telecomm Interestedness services Mail address 14< Stat.Abs.U.S.91 1,16E-04 Phone dialling 10< FCC Stat.Com.Carr. 5,49E-08 A.3 Financial services Interestedness, law ATM transactions 11< Func.cost. anal.199 3,62E-07 Applications for loans 12< Natl.avg.rept, 1990 Checking account 12< Natl.avg.rept, 1990 2,30E-07 transactions Savings account 12< Natl.avg.rept, 1990 transactions A.4 Medical services Interestedness, law B. Copies of contracts Interestedness and agreements C. Information on Law delivery of goods and services Retail bills, receipts 13< Stat.abs.U.S.91 2,49E-05 D. Information provided to consumers Notice of usage Law Contents information on Law 14< Stat.abs.U.S.91 3,70E-04 food products E. Payment information Interestedness, law F. Information on Interestedness, law Stat.abs.U.S.91 soliciting job G. Information provided for courts By ensuer in civil Interestedness, law 20< lawsuits By subpoenad witness Law 20< 2,74E-02 H. Information provided Law to authorities For a driver’s license Law For recording birth Law 8< Stat.abs.U.S.91 2,90E-10 For getting permission to Conduct profession Law Produce drugs Law Distribute films Law

Chapter 5. New research in Berkeley

The new research object - transplanted from Hungary to Berkeley - has been soon adopted and got several followers, after the two-semester Fulbright term of the present author had ended. Once School of Library and Information Sciences has got a new mission which is reflected in the new name: School of Information and Management. Professor Varian a world-wide known information-economist - has been appointed to the Dean of the school. For the time being, five persons conduct a world-wide survey at UC Berkeley, also publishing data for the U.S. The Berkeley group focused its attention to some of the SSNIA indicators They redefined several media and estimate yearly „information production”, „information consumption”, „accumulated stocks of information”, rates of growth and „other valuables of interest”. "Produced", i.e. outputted, commodities sometimes are called "flow" as a contrary to "stocks".

Senior researchers Peter Lyman and Hal Varian [Varian00] have published the results of the study entitled „How much information?” [www.berkeley.edu]

Any measurement which aspires to the rank of being qualified as „technical” or „scientific”, should have a well founded and well defined methodology. Information – the fundamental concept of classical theory of communications – assumes a transmitter and a receiver and a noisy telecommunication channel. There is no such amount there, as an immanent, absolute quantity of information „in a book” and as a thing as “production or consumption of information” in classic theory. [Braman] in her clasical study collected several definitions to „information”, none of which is suitable for measuring „information” in an economic context. Furthermore, production, output, flow, consumption are internationally standardised concepts and quantities which are measured in the frames of theory and methodology of national accounting [SNA]. In their recent study, Lyman and Varian do not provide a new generic definition of their own, concerning their variables-indicators. Unless one created a new definition, he should decide whether he - as an economist - is interested in the macroeconomy of production/consumption/accumulation of goods and services or, as an electric engineer is interested in transmission of signals in noisy channels. The elements of the two theories must not be mixed in some arbitrary ways. The authors mistakenly adopted the information paradigm from the theory of communications and tried to merge it into macroeconomics.

The authors adopt a specialized “store model”. As the authors write: “Soon it will be technologically possible for an average person to access virtually all recorded information. The natural question then becomes: How much information is there to store?”

This approach is straightforward for a company, like the storage media company which financed the study, or someone, who has a vision of a future information society with monopolistic content distributor. However, for a citizen/netizen or policy maker, this is not an issue. A store modell like this, has nothing in common with political economy of information commodities. A policy maker should be interested in economic issues. The „production” and „consumption” figures of the authors can not be directly matched with "value" figures. Major policy issues in economy are related to rarity, uneven distribution, local lack of commodities. Millions of people just can not access the Net or can download only insignificant volumes from the surface and never have production tools which make them to be able to profit from those remarkable terabytes. In a country, like Hungary, telecommunications companies may provide services on prices which are higher than those in the developed countries. Millions never can transgress “media information”, the kind of the newest ambushes in Palestine, computer games, pop stars – the old content in a new digital form.

The study do not deals with us, human individuals, whose information storage and processing capability, with some1012-14 neural cells, and 103 links each, is still much more than anything mentioned in Varian's study. Humans should stand in the centre of any serious worldwide quantitative descriptive economic model and theory of information society.

One of the conclusions of the study that while “information production” is growing exponentially, “information consumption” actually does not change. This is not surprising, given Lyman and Varian are neglecting “machine information consumption”, i.e. consumption of information commodities by owners of machines. This is consumption of commodities - like floppies, on-line signals, etc. - they input into their machines. Machines actually receive more and more information, which is the far most important process of information economy and policy. The more computers are installed, the more Winchesters, CD-s and floppies are not only produced, and consumed, but equally importantly whose contents are consumed. Ignoring consumption of floppies and Winchesters when accounting their production is just misleading. Statistics regularly account "consumption by one's machines" as one's consumption. Examples can be "fuel consumption by cars", "electric energy consumption by household equipment, or factory equipment".

Theoretical and methodological sloppiness, dull conceptual background is most outstanding in the chapter on broadcast media. Broadcast media are "par excellence" mass media. Surprisingly enough the figures of production of mass communication are based upon broadcast time only, and the authors do not take into account the billions of hours millions of viewers/listeners with watching/listening. A mass medium without its audience, daily reach, would not be a mass medium at all. Actually, since communication of stations with their daily reach is accounted as if it were one channel with one viewer, it is not surprising, that phone conversations were found to be more productive than mass media. Furthermore, the authors are not sure whether radio broadcasting and print media should be measured in compressed or uncompressed bytes. For a statistician it is clear that statistics should reflect facts. It is a matter of fact that broadcasting stations transmit programs, and never compressed programs.

Methodological inadequacy and sloppines have lead the authors to the assumption of „democratisation of „data””. In fact nothing like democratization can be concluded from Varian's data. "Data" can not be "democratic" or "more democratic". It is a society, which can be - more or less - democratic. While a vast amount of unique information is actually created and stored by individuals, this does not necessarily imply, that they can play a more significant role in political decision making. The right for a professional to sit behind a screen, and gaze at EXCEL tables for eight hours daily, is probably much more blue-collarization of white-collars than democratization of society. Most data on personal computers can never serve as capital, having neither commercial value nor the capability of providing capital income to the owner, even if millions have acquire fortunes by building their busines by PC-s and on the Internet.

Chapter 6. A European Initiative - ESIA

In the frames of a co-operation of Dutch, Austrian, Hungarian, Italian and Spanish researchers an effort has been made to define ESIA - an European Standard System of Information Accounts.

Chapter 7. SSNIA Version 1.2

The rapid development of technology has allowed that new statistical procedures would be defined at the human/machine and machine/machine interfaces. Since 1994 a number of standard indicators have been introduced to measure information flow on electronic networks. Significant development of cognitive psychology, statistical pychology and neurolingusitics have allowed to refine models of human brain and statistical procedures to measure quantitative changes.

2. Strength and weaknesses of various efforts

Natural-unit terms and value-unit terms reflect different aspects of transactions. SSNIA's solution to a dual account of phenomena and processes promises more than the one-sided approaches. [Pool] and [Varian00]'s "unitary costs" reflect only a particular moment of value flows escorting the flows of information commodities.

Table 9. Units of measurements in the accounts

No Unit of measurement de Sola Pool SSNIA 1.0 Included in [Varian] 1 word yes no no 2 bit no yes yes 3 monetary unit no yes no

The sophisticated system of SNA indicators have been developed for following the production, distribution and redistribution of commodities. Production, reproduction, distribution and redistribution of information commodities can not be followed unless the whole system of SNA indicators were applied. For instance, the output of books and their consumption shows a dynamic growth in value terms in Hungary in the 1990-2000 period. However, in natural unit terms, both output and consumption show a stagnation or a decrease in the same period. When the accumulation of human knowledge - from books - are studied, it is obvious, that the growth of consumption in value terms is due to the increasing of profit margins, production and material cost. No more, but less information - from books - has been accumulated by human individuals - something, which is hidden by the figures in value terms. More color pictures, better quality, however, favors to easier reading.

Table 10. Important indicators in various studies No Indicator de Sola Pool SSNIA [Varian] 1 production yes In accordance with SNA, production yes is not an indicator, but a process 2 output no yes no 3 consumption yes yes mostly no 4 stock no yes yes 5 accumulation no yes no 6 exports no yes no 7 imports no yes no 8 externalities no yes no 9 input no yes no 10 use no yes no 11 information added no yes no 12 transfer no yes no 13 asset no yes no 14 valuable no yes no 15 capital no yes no 16 depreciation no yes no Any serious account should have a grasp, big enough to cover a reasonable majority of actors, transactions, commodities.

If several small actors are defined, the output of services will be significantly greater, than in the case, when big units are defined as actors. If no more than one actor is defined, the amount of output should be zero, since any output assumes a consumer, to whom the output is outputted. Without defining classes of actors no serious accounts can be prepared.

Table11. Classes of commodities in various studies

No Commodity class de Sola Pool SSNIA 1.0 [Varian] 1 Radio Programming and S/M/L and stereo; Programming only broadcasting programming, broadcasting and presentation are accounted 2 TV B/w, colour programming Programming, broadcasting Programming only and broadcasting and presentation are accounted 3 Cable TV Programming and Programming, broadcasting Programming only broadcasting and presentation are accounted 4 Record, magnetic tape Yes LP, MC records, 4,5 mm Yes tapes, recorded and unrecorded cassettes 5 Movies Yes Feature films, No documentaries, 8 and 16mm films 6 Education Yes primary, secondary and No tertiary education 7 Newspaper Yes Yes Yes 8 Magazine Yes Yes Ye 9 Book Ye Ye Yes 10 Telephone directory Yes Included in books No 11 Direct mail Yes Included in mails No 12 Mail Yes Yes Yes 13 Phone Yes Local, domestic and Yes international 14 Telex Yes Yes No 15 Telegraph Yes Voice and paper delivery No 16 Mailgram No No No 17 Fax Yes Yes No 18 Data Yes Yes No 19 Office documents No Yes Yes 20 www No No Yes 21 e-mail, mailing list No Yes Yes 22 Usenet No No Yes 23 FTP No No Yes 24 IRC, messaging services, No No Yes Telnet 25 Hard disk drives No Yes Yes 26 Floppy disks No Yes Yes 27 Removable magnetic disk No Yes, included in hard disk Yes drives drives 28 Digital data creation No Yes: "keypunching" Yes 29 Optical disks (CD, CD-ROM) No Ye Yes 30 Photofilms No Yes Yes, paper-based photoes included 31 Motion pictures No Ye Yes 32 X-ray films No Yes Yes 33 Recorded ROM, RAM, chip- No Yes No memories 34 Personal communications No Yes No 35 Human knowledge Bo Yes No 36 TV show and upplying radio No Yes No program 37 R+D, spectator sports No Yes No institutions, theatres, musei, culture and entertainment

Similarly to SSNIA, [Varian00]'s interest extends both to the "originals" and "copies". The definition of "originals" sometimes is not trivial, or even if the definition would be clear, no statistics can be collected. For instance, majority of radio programmes is nothing else then broadcasting of canned music, replay of records, accounted as such there. For several countries, no detailed statistics are available on the volume of canned and alive programmes. On the other hand, total broadcast time can not be considered - like Varian does - the volume of information produced by program-makers and broadcasters. .Broadcast time characterizes the volume of programming, and the sum of all receiver- hours of all listeners characterizes the performance of broadcasters.

There are several transactions related to the web technology, of which Varian's account extends to the accumulation of contents only. He does not even consider the output and consumption of web services, when copies are reproduced for the user to be displayed on his screen, even though several figures are available for their volume (sessions, visits, page-impressions, click-ons or downloads of the pages).

Table 12. Classes of actors

No Actor class de Sola Pool SSNIA 1.0 [Varian] 1 governments no yes no 2 individuals no yes no 3 enterprises no yes no 4 non-profits no yes no 5 nations yes yes yes 6 ethnic and language communities no yes no 7 regions no yes no

Table 13. Classes of classifications

No Classification de Sola Pool SSNIA 1.0 [Varian] 1 Actors no yes no 2 Transactions no yes no 3 Commodities yes yes yes 4 Sectors no yes no 5 Industries no yes yes 6 Stocks no yes no

There are some commodities, whose definitions in the different studies are similar enough to compare their bit-eqivalents.

The differences in the values of the bit-equivalents may reflect differences in the definitions in various systems. A hypothetic scanning as a tool to define the volume of information provides significantly higher figures, than character-wise re-keypunching. [SNIA92] makes distinction between sensory (scanning) and substantial amount/volume of information carried by a non-digital information commodity.

Compression is another important factor. Each and any file is either compressed or non-compressed. Statistics is a science about measuring, observing reality, „as is”. Though the bit-volume of all files recorded intentionally compressed or not on a digital carrier is a factual number, which is known for users, [Varian00] adopts an additional compressed volume.

The differences may reflect real differences or different assumptions concerning the nature of the commodity. While definition of "print sheets in a book" is an international definition, the average number of figures and pictures, particularly colored pictures, in books may be different in various countries in various years. The issues of newspapers are much more voluminous in the U.S., than in Hungary, due to excess ad-attachments and thematic enclosures, their bit-equivalents must be different.

The great differences in Table 5. can be explained by the differences of definitions and assumptions that various authors introduced. USSNIA assumed hypothetic „scanning”, [Varian00] hypothetic scanning and re-keypunching to determine volume of information in print media. [Varian00] assumed figures, which differ from figures for standard channels assumed in USSNIA. This shows how necessary the international standardization of information accounting is.

Due to technical progress, the values of equivalents prepared for different years, are obviously different. The average size of a Winchester in circulation in 1990 assumably did not exceed 100 Mbyte, while it is well over 20 Gbyte in 2002. Technical progress penetrates rich countries first: the average size of Winchesters in the rich and technically more developed U.S. is greater than it is in poor countries in the same year. Table 14. Bit-equivalents of some comparable commodities in USSNIA and [Varian00]

No Commodity class USSNIA 1994 Varian 1999 1 Radio broadcasting: hour 0,63 Gb/hour 0,05 Gb/hour 2 TV broadcasting: hour 46,6 Gb/hour 1,3 Gb/hour 3 Audiocassette 0,95 Gb/tape 1 Gb/tape (Blank audio tape) 4 Videotapes 184/Gb/tape 4 Gb/tape (Blank video tape) 5 Newspaper 283 Gb/year/title 115 Mb/year/title 6 Magazine 9,6 Gb/year/title 225-650 Mb/year/title 7 Book 4,36 Gb/book 41 Mb/book 8 Mail 120 Kb/mail 75 Kb/mail 9 Phone 1,32 Mb/DAM 0,96 Mb/perc 10 Floppy disks 1,44Mb/disk 1,44 Mb/disk

3. Future: The floor is now open

Ten years ago, when SNIA and SSNIA were invented, the situation was premature to launch a concerted international effort to define and introduce a SNIA into the system of standard official statistics.

Meanwhile vast majority of population in the developed countries have become regular users, several times the slaves, of digital information technology. For millions, the fact that all kinds of human and non-human information are convertible, now is obvious. Professionals of nations need and demand real figures which reflect information activity of their own, governments, enterprises, various big brothers, real economic flows and knowledge stocks, standard figures by official statistical institutions. Traditional macroeconomic tables do not cover basic processes and phenomena in information societies. Primitive eclectic "rankings" of nations can not replace statistics, which show undergoing deep macroeconomic processes, "the essence" of information societies. While the EU, and the OECD urge an "information society", information commodities have either not been defined. Copyright and intellectual property are key issues, but their treatment in accounts of the official statistics is not yet well solved. The efforts of OECD and EUROSTAT to define ICT sectors of the economy, and ICT commodities regularly neglect human information and non-digital world - the majority.

The recent discovery of correlation between nation's GDP and population's average IQ [Lynn02], underscores the importance of serious information statistics. Time is ripe for an international scientific cooperation of researchers in statistical offices and educational organisations, devoted to seeking of truth and welfare of nations, and also independent enough from various power groups, can lead to really useful information statistics, to the implementation of a system of standard information accounts, as contoured in SSNIA 1.0 and 2.0.

Its time scientific community and community of official statisticians joined and defined information age accounts for information societies. This might prevent laymen and digital prophets to blow their trumpets and allows citizens to take part in the very creation of new societies. This will be the real "democratisation of data". A common language concerning the core macroeconomic issues in information societies is a core issue in international understanding. The Fulbright Program with his special tools might promote the process.

References

[Braman90] Braman S. (1990) Defining Information: An Approach for Policymakers Telecommunications Policy 13(3) pp 233-42

[DSP] De Sola Pool I.et al. (1984): Communications Flows. A Census in the United States and Japan. North Holland, Amsterdam, New York, Oxford, University of Tokyo Press

[Dienes86] Dienes István (1986): Magnitudes of the Knowledge Stock and Information Flows in the Hungarian Economy (in Hungarian) In. Tanulmányok az információgazdaságról KSH-OMIKK, 1996, Bp. pp. 89-101.

[Dienes93] Dienes István (1993): Towards a System of National Information Accounts Proceedings of the XXIst Telecommunications Policy Research Conference, Solomons, Maryland http://free.x3.hu/infostat

[Dienes92] Dienes István (1992): Information balances of Hungary 1975-90 Presentation at the Conference "Information economy, information policy", Budapest, 1992 http://free.x3.hu/infostat

[Dienes94] Dienes István (1994): Accounting the information flows and knowledge stocks in the U.S.: Preliminary results. Presentation in Berkeley http://www.geocities.com/gsnmrjoeeye/Mypage

[Dienes96] Conference of the IARIW, New-Delhi

[Dienes95] SYSTEM OF NATIONAL INFORMATION ACCOUNTS of The United States of America, 1970--1990 http://www.geocities.com/gsnmrjoeeye/Mypage

[Dienes97] http://free.x3.hu/infostat

[ESIA] F. Bellini, Y. Braunstein , G.M. Clavero, M. Deistler, I. Dienes, T. Jellema, Á. Kozák, I. Paterson, E. Skriner, , (2001) European Standard System of Information Accounts Project proposal of research consortium, submitted to the 5th Programme of the EU

[Dienes00] Dienes István (2000). What is how much? Misleading figures from Berkeley http://free.x3.hu/infostat

[Dienes02] Dienes István (2002) National Accounting of Information. Reference Manual of SSNIA, Version 1.1 Manuscript, Budapest.

[Lakoff] G. Lakoff (1987): Women, Fire and Dangerous Things. What Categories Reveal about the Mind. Chicago, The University of Chicago Press

[Lesk] Michael Lesk (1997): How much information Is there in the world? http://www.lesk.com/mlesk/ksg97/ksg.html

[Lynn02] Richard Lynn, Tatu Vanhanen (2002): IQ and the Wealth of Nations. Praeger, Westport, CT. 256 pp.

[Odlyzko] Andrew Odlyzko (2000): Content is not the king Technical Rept., AT&T Labs http://www.research.att.com/~amo/doc/networks.html

[Repo] Aatto Repo (1987): Espo, Techn. Rept.

[SNA] System of National Accounts (1993) Inter-Secretariat Working Group on National Accounts, Eurostat, IMF, OECD, UNSO, World Bank.

[SNIA] Dienes I. (1994) National Accounting of Information. Reference Manual of SNIA, Version 1.1 Manuscript, Berkeley, Budapest http://free.x3.hu/infostat/ [Varian99] Varian, Hal, Lyman Coopers: (1999): How much information? http://www.sims.berkeley.edu/ How much information? Accounting the information flows and knowledge stocks István Dienes - scientific advisor Cég-INFO Ltd., Budapest [email protected], http://free.x3.hu/infostat Table 1. Units of measurements in the accounts

No Unit of measurement de Sola Pool SSNIA 1.0 Included in [Varian] 1 word yes no no 2 bit no yes yes 3 monetary unit no yes no

Table 2. Important indicators in various studies

No Indicator de Sola Pool SSNIA [Varian] 1 production yes In accordance with SNA, yes production is not an indicator, but a process 2 output no yes no 3 consumption yes yes mostly no 4 stock no yes yes 5 accumulation no yes no 6 exports no yes no 7 imports no yes no 8 externalities no yes no 9 input no yes no 10 use no yes no 11 information added no yes no 12 transfer no yes no 13 asset no yes no 14 valuable no yes no 15 capital no yes no 16 depreciation no yes no Table 3. Classes of commodities in various studies

Table 5. Classes of classifications

No Classification de Sola Pool SSNIA 1.0 [Varian] 1 Actors no yes no 2 Transactions no yes no 3 Commodities yes yes yes 4 Sectors no yes no 5 Industries no yes yes 6 Stocks no yes no

Table 6. Bit-equivalents of some comparable commodities in USSNIA and [Varian00]